|Titre descriptif du dossier proposé :||A chatbot to help students answer common IT questions any time, anywhere|
|Année de la candidature||2019|
|Awarded||National prize - Innovation|
|Nom et titre de l’auteur (des auteurs) de l’idée||
|Nom de l’établissement :||University of Alberta|
|Adresse au bureau :||3-02B General Services Building
University of Alberta
Edmonton, Alberta T6G 2H1
|Numéros de téléphone||(780) 492-3573|
|Nom (Responsable en chef des services administratifs de l’établissement)||Gitta Kulczycki|
|Titre (Responsable en chef des services administratifs de l’établissement)||Vice-President (Finance & Administration)|
|Office Address||2-04B South Academic Building
University of Alberta
Edmonton, Alberta T6G 2J7
|Numéros de téléphone||(780) 492-8254|
Eve is an artificial intelligence engine designed to answer IT-related questions over live chat. With Eve on the team, IST’s live chat is available 24/7 instead of just during defined hours, so students and staff can get IT help around the clock.
Eve offers IST clients a shorter wait time and a quicker response to common questions and issues, such as how to connect to wireless, where to find coffee and where to find free software. In fact, she can answer over 100 types of question - and more are added weekly.
With Eve helping IST clients over live chat, our Service Desk teammates are able to focus on larger and more complex tasks.
In September alone Eve engaged in over 1,000 live chats. When she was stumped, she transferred to a human 18 times and created a ticket for a human to review later 23 times, which can also happen after business hours.
Eve is built on IBM's Watson AI platform, which means any institution could do the same thing for minimal cost - our bill from IBM is generally less than $10 per month. While the answers are unique to each institution, the questions will be similar.
Eve continues to evolve. New features are continually rolling out and we are looking to expand Eve's knowledge to include general student questions and help people through the applications process.
|Critères||Please submit one paragraph describing how the proposal fulfills each of the evaluation criteria.|
The problems we are solving are general to every institution:
Every institution has units and departments with frequently asked questions, whether from staff or students - indeed, most of those questions are also common across institutions.
It’s also true that most institutions have constituents in many time zones, making office-hours support problematic for many.
Additionally, all institutions have periods of peak demand, generally many times greater than the average, and must either “staff up” for peak periods or allow service levels to suffer.
Eve is maintained by two junior analysts part-time, not expensive AI specialists. IBM Watson's pricing is negligible, making it accessible to even institutions on the tightest of budgets.
A low-cost solution, requiring easily obtained skills, and solving problems found in all institutions make this highly transferable.
|Impact sur la qualité||
Eve provides consistency of information when giving an answer. Although Service Desk agents use standard knowledge base articles they often use experience to assist the user which can produce higher quality assistance but also introduces nuances which mean the experience of a user can vary depending on the analyst they speak with. Eve will give the same accurate and factual answer each time.
To restate: Eve is essentially infinitely scalable for a trivial cost: every student could ask a question at the same time and there would be no wait time. The September rush does not affect IST’s ability to deliver service through Eve at all.
|Impact sur la productivité||
The intent was to extend service beyond business hours and to provide faster service at peak times. We have managed to extend service to 24/7 with a minimal increase in costs (Watson costs less than $10/month on average).
Approximately 200 chats per month are handled by Eve, spiking to over 1,000 in September. The average internal cost of a Service Desk transaction via chat or phone is $10 which equates to $2000 a month in savings of service desk analyst time, or $10,580 in September.
As we move out of IT and into other areas of the institutions, imagine providing immediate answers to a broad range of applicants’ questions at 1am, during peak period - in their own language!
Higher education has always struggled to provide timely responses to applicants, students and staff at peak times, or outside of business hours when they are in a very different time zone. Introducing a chatbot provides a cost-effective answer to a real and universal problem. By using IBM's artificial intelligence platform we are tapping into some of the world's best natural language processing and it will continue to improve over time due to the platform it's on. Eve provides immediate answers to the most common questions, from the devices students prefer to use, when and where they have the question.
Centre du savoir
Dernières nouvelles / Sujets de l'heure